| Message ID | 20200316111034.1686721-1-numans@ovn.org |
|---|---|
| Headers | show |
| Series | Incremental processing improvements. | expand |
On Mon, Mar 16, 2020 at 4:11 AM <numans@ovn.org> wrote: > > From: Numan Siddique <numans@ovn.org> > > (Removed the RFC tag and Submitting these patches as formal patches now) > > This patch series tries to handle port binding, datapath binding > and ovs interface changes incrementally. > > Below are the results of some testing I did with ovn-fake-multinode > setup > > Test setup > ------ > 1. ovn-central fake node running OVN dbs and 2 compute nodes running > ovn-controller. > > 2. Before running the tests, used an existing OVN db with the below > resources > No of logical switches - 53 > No of logical ports - 1256 > No of logical routers - 9 > No of logical router ports - 56 > No of port groups - 152 > No of logical flows - 45447 > > Port bindings on compute-1 - 19 > Port bindings on compute-2 - 18 > No of OF flows on compute-1 - 84996 > No of OF flows on compute-2 - 84901 > > 3. The test does the following > - Creates 2 logical switches (one for each compute node) and connect to a > logical router for each compute node. > - 100 logical ports are created (50 per lswitch), a simple ACL is added and the address > set is created for each port. > - Each port is bound on the respective compute node and the test > pings the IP of the port (from another port belonging to the same > lswitch created earlier). > > > Below are the results with OVN master > > +----------------------------------------------------------------------------------------------------------+ > | Response Times (sec) | > +----------------------------------+--------+--------+--------+--------+--------+--------+---------+-------+ > | action | min | median | 90%ile | 95%ile | max | avg | success | count | > +----------------------------------+--------+--------+--------+--------+--------+--------+---------+-------+ > | ovn.create_or_update_address_set | 0.484 | 0.507 | 0.526 | 0.537 | 0.547 | 0.51 | 100.0% | 100 | > | ovn.create_port_acls | 0.948 | 1.014 | 1.05 | 1.06 | 1.089 | 1.016 | 100.0% | 100 | > | ovn_network.bind_port | 1.244 | 1.294 | 1.333 | 1.346 | 1.397 | 1.301 | 100.0% | 100 | > | ovn.bind_ovs_vm | 0.379 | 0.432 | 0.463 | 0.469 | 0.472 | 0.434 | 100.0% | 100 | > | ovn.bind_internal_vm | 0.824 | 0.861 | 0.897 | 0.906 | 0.941 | 0.868 | 100.0% | 100 | > | ovn_network.wait_port_ping | 7.153 | 7.183 | 7.214 | 7.22 | 7.237 | 7.185 | 100.0% | 100 | > | total | 10.373 | 10.479 | 10.544 | 10.551 | 10.581 | 10.483 | 100.0% | 100 | > +----------------------------------+--------+--------+--------+--------+--------+--------+---------+-------+ > Load duration: 1048.8911039829254 > Full duration: 1050.4055325984955 > > Below are the results with these patches > > +-------------------------------------------------------------------------------------------------------+ > | Response Times (sec) | > +----------------------------------+-------+--------+--------+--------+-------+-------+---------+-------+ > | action | min | median | 90%ile | 95%ile | max | avg | success | count | > +----------------------------------+-------+--------+--------+--------+-------+-------+---------+-------+ > | ovn.create_or_update_address_set | 0.479 | 0.492 | 0.513 | 0.52 | 0.531 | 0.495 | 100.0% | 100 | > | ovn.create_port_acls | 0.946 | 0.964 | 0.998 | 1.001 | 1.041 | 0.969 | 100.0% | 100 | > | ovn_network.bind_port | 1.22 | 1.297 | 1.342 | 1.36 | 1.402 | 1.298 | 100.0% | 100 | > | ovn.bind_ovs_vm | 0.373 | 0.418 | 0.436 | 0.455 | 0.472 | 0.42 | 100.0% | 100 | > | ovn.bind_internal_vm | 0.818 | 0.88 | 0.911 | 0.93 | 0.982 | 0.878 | 100.0% | 100 | > | ovn_network.wait_port_ping | 3.812 | 3.854 | 3.906 | 3.934 | 4.015 | 3.861 | 100.0% | 100 | > | total | 6.977 | 7.165 | 7.277 | 7.308 | 7.401 | 7.176 | 100.0% | 100 | > +----------------------------------+-------+--------+--------+--------+-------+-------+---------+-------+ > Load duration: 718.401807308197 > Full duration: 719.8932237625122 > > > Testing the same with fresh OVN dbs didn't show any improvements in the > timings. > > Hi Numan, Thanks for taking this hard rock. The result looks promising! I haven't finished reviewing all the patches yet, but I have some comments that I think are worth discussing as early as possible, so I send them for now. Please find them in my reply for patch 1, 2, 3 and 6. Thanks, Han > RFC v2 -> v1 > --------- > * Fixed the 2 failing test cases. > * Updated the commit messages. > > > RFC v1 -> RFC v2 > --------- > * Added 2 new patches > * Patch 5 (ofctrl_check_and_add_flow) was submitted earlier too and > the previous discussion is here - https://patchwork.ozlabs.org/patch/1202417/ > * Patch 6 handles I-P for ct_zone and OVS interface changes in > flow_output_run stage. > Numan Siddique (6): > Refactor binding_run()to take two context argument - binding_ctx_in > and binding_ctx_out. > ovn-controller: Refactor binding.c > ovn-controller: I-P for port binding in runtime_data stage > ovn-controller: I-P for datapath binding > ofctrl_check_and_add_flow: Replace the actions of an existing flow if > actions have changed. > ovn-controller: I-P for ct zone and OV > > > Numan Siddique (6): > Refactor binding_run()to take two context argument - binding_ctx_in > and binding_ctx_out. > ovn-controller: Store the local port bindings in the runtime data I-P > state. > ovn-controller: I-P for SB port binding and OVS interface in > runtime_data. > ovn-controller: I-P for datapath binding > ofctrl_check_and_add_flow: Replace the actions of an existing flow if > actions have changed. > ovn-controller: I-P for ct zone and OVS interface changes in flow > output stage. > > controller/binding.c | 1112 ++++++++++++++++++++++++----------- > controller/binding.h | 58 +- > controller/lflow.c | 13 + > controller/lflow.h | 2 + > controller/ofctrl.c | 23 +- > controller/ovn-controller.c | 332 ++++++++--- > controller/ovn-controller.h | 22 + > controller/physical.c | 48 ++ > controller/physical.h | 5 +- > controller/pinctrl.c | 19 +- > controller/pinctrl.h | 4 +- > tests/ovn-performance.at | 22 +- > 12 files changed, 1197 insertions(+), 463 deletions(-) > > -- > 2.24.1 > > _______________________________________________ > dev mailing list > dev@openvswitch.org > https://mail.openvswitch.org/mailman/listinfo/ovs-dev
From: Numan Siddique <numans@ovn.org> (Removed the RFC tag and Submitting these patches as formal patches now) This patch series tries to handle port binding, datapath binding and ovs interface changes incrementally. Below are the results of some testing I did with ovn-fake-multinode setup Test setup ------ 1. ovn-central fake node running OVN dbs and 2 compute nodes running ovn-controller. 2. Before running the tests, used an existing OVN db with the below resources No of logical switches - 53 No of logical ports - 1256 No of logical routers - 9 No of logical router ports - 56 No of port groups - 152 No of logical flows - 45447 Port bindings on compute-1 - 19 Port bindings on compute-2 - 18 No of OF flows on compute-1 - 84996 No of OF flows on compute-2 - 84901 3. The test does the following - Creates 2 logical switches (one for each compute node) and connect to a logical router for each compute node. - 100 logical ports are created (50 per lswitch), a simple ACL is added and the address set is created for each port. - Each port is bound on the respective compute node and the test pings the IP of the port (from another port belonging to the same lswitch created earlier). Below are the results with OVN master +----------------------------------------------------------------------------------------------------------+ | Response Times (sec) | +----------------------------------+--------+--------+--------+--------+--------+--------+---------+-------+ | action | min | median | 90%ile | 95%ile | max | avg | success | count | +----------------------------------+--------+--------+--------+--------+--------+--------+---------+-------+ | ovn.create_or_update_address_set | 0.484 | 0.507 | 0.526 | 0.537 | 0.547 | 0.51 | 100.0% | 100 | | ovn.create_port_acls | 0.948 | 1.014 | 1.05 | 1.06 | 1.089 | 1.016 | 100.0% | 100 | | ovn_network.bind_port | 1.244 | 1.294 | 1.333 | 1.346 | 1.397 | 1.301 | 100.0% | 100 | | ovn.bind_ovs_vm | 0.379 | 0.432 | 0.463 | 0.469 | 0.472 | 0.434 | 100.0% | 100 | | ovn.bind_internal_vm | 0.824 | 0.861 | 0.897 | 0.906 | 0.941 | 0.868 | 100.0% | 100 | | ovn_network.wait_port_ping | 7.153 | 7.183 | 7.214 | 7.22 | 7.237 | 7.185 | 100.0% | 100 | | total | 10.373 | 10.479 | 10.544 | 10.551 | 10.581 | 10.483 | 100.0% | 100 | +----------------------------------+--------+--------+--------+--------+--------+--------+---------+-------+ Load duration: 1048.8911039829254 Full duration: 1050.4055325984955 Below are the results with these patches +-------------------------------------------------------------------------------------------------------+ | Response Times (sec) | +----------------------------------+-------+--------+--------+--------+-------+-------+---------+-------+ | action | min | median | 90%ile | 95%ile | max | avg | success | count | +----------------------------------+-------+--------+--------+--------+-------+-------+---------+-------+ | ovn.create_or_update_address_set | 0.479 | 0.492 | 0.513 | 0.52 | 0.531 | 0.495 | 100.0% | 100 | | ovn.create_port_acls | 0.946 | 0.964 | 0.998 | 1.001 | 1.041 | 0.969 | 100.0% | 100 | | ovn_network.bind_port | 1.22 | 1.297 | 1.342 | 1.36 | 1.402 | 1.298 | 100.0% | 100 | | ovn.bind_ovs_vm | 0.373 | 0.418 | 0.436 | 0.455 | 0.472 | 0.42 | 100.0% | 100 | | ovn.bind_internal_vm | 0.818 | 0.88 | 0.911 | 0.93 | 0.982 | 0.878 | 100.0% | 100 | | ovn_network.wait_port_ping | 3.812 | 3.854 | 3.906 | 3.934 | 4.015 | 3.861 | 100.0% | 100 | | total | 6.977 | 7.165 | 7.277 | 7.308 | 7.401 | 7.176 | 100.0% | 100 | +----------------------------------+-------+--------+--------+--------+-------+-------+---------+-------+ Load duration: 718.401807308197 Full duration: 719.8932237625122 Testing the same with fresh OVN dbs didn't show any improvements in the timings. RFC v2 -> v1 --------- * Fixed the 2 failing test cases. * Updated the commit messages. RFC v1 -> RFC v2 --------- * Added 2 new patches * Patch 5 (ofctrl_check_and_add_flow) was submitted earlier too and the previous discussion is here - https://patchwork.ozlabs.org/patch/1202417/ * Patch 6 handles I-P for ct_zone and OVS interface changes in flow_output_run stage. Numan Siddique (6): Refactor binding_run()to take two context argument - binding_ctx_in and binding_ctx_out. ovn-controller: Refactor binding.c ovn-controller: I-P for port binding in runtime_data stage ovn-controller: I-P for datapath binding ofctrl_check_and_add_flow: Replace the actions of an existing flow if actions have changed. ovn-controller: I-P for ct zone and OV Numan Siddique (6): Refactor binding_run()to take two context argument - binding_ctx_in and binding_ctx_out. ovn-controller: Store the local port bindings in the runtime data I-P state. ovn-controller: I-P for SB port binding and OVS interface in runtime_data. ovn-controller: I-P for datapath binding ofctrl_check_and_add_flow: Replace the actions of an existing flow if actions have changed. ovn-controller: I-P for ct zone and OVS interface changes in flow output stage. controller/binding.c | 1112 ++++++++++++++++++++++++----------- controller/binding.h | 58 +- controller/lflow.c | 13 + controller/lflow.h | 2 + controller/ofctrl.c | 23 +- controller/ovn-controller.c | 332 ++++++++--- controller/ovn-controller.h | 22 + controller/physical.c | 48 ++ controller/physical.h | 5 +- controller/pinctrl.c | 19 +- controller/pinctrl.h | 4 +- tests/ovn-performance.at | 22 +- 12 files changed, 1197 insertions(+), 463 deletions(-)